bioRxiv preprint doi: https://doi.org/10.1101/819136; this version posted October 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. A specialized deceptive pollination system based on elaborate mushroom mimicry Satoshi Kakishima1, Nobuko Tuno2, Kentaro Hosaka1, Tomoko Okamoto3, Takuro Ito1, and Yudai Okuyama1* 1. Tsukuba Botanical Garden, National Museum of Nature and Science, Amakubo 4-1-1, Tsukuba, Ibaraki, 305- 0005, Japan 2. Laboratory of Ecology, Graduate School of Natural Science and Technology, Kanazawa University, Kanazawa 920-1192, Japan. 3. Laboratory of Insect Ecology, Faculty of Applied Biological Sciences, Gifu University, Yanagido 1-1, Gifu 501- 1193, Japan *Author for correspondence: Yudai Okuyama e-mail: [email protected] phone: +81-29-853-8435 bioRxiv preprint doi: https://doi.org/10.1101/819136; this version posted October 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. Abstract Despite its potential effectiveness for outcrossing, few examples of pollination via mushroom mimicry have been reported. This may be because the conditions under which the strategy can evolve are limited and/or because demonstrating it is challenging. Arisaema is a plant genus that has been suggested to adopt mushroom mimicry for pollination, although no compelling evidence for this has yet been demonstrated. Here, we report that Arisaema sikokianum utilizes mostly a single genus of obligate mycophagous flies (Mycodrosophila) as pollinators, and that the insect community dominated by Mycodrosophila is strikingly similar to those found on some species of wood-decaying fungi. Comparative chemical analyses of Arisaema spp. and various mushrooms further revealed that only A. sikokianum emits a set of volatile compounds shared with some mushroom species utilized by Mycodrosophila. Meanwhile, other closely related and often sympatric Arisaema species do not possess such typical traits of mushroom mimicry or attract Mycodrosophila, thereby likely achieving substantial reproductive isolation from A. sikokianum. Our finding indicates that mushroom mimicry is an exceptional and derived state in the genus Arisaema, thus providing an unprecedented opportunity to study the mechanisms underlying the coordinated acquisition of mimicry traits that occurred during a recent speciation event. Keywords: Arisaema, deceptive pollination, Drosophilidae, floral scent, floral mimicry, mushroom bioRxiv preprint doi: https://doi.org/10.1101/819136; this version posted October 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. Introduction The diversity of floral forms found in angiosperms reflects the diverse plant–pollinator interactions that flowers engage in, wherein a particular suite of floral traits adapted to a specific functional group of pollinators is termed ‘pollination syndrome’ (Fenster et al. 2000, Ollerton et al. 2009). Among such relationships, a non-negligible proportion (associated with 7500 plant spp.) are based on deceptive pollination strategies (Renner 2006), in which plants lure pollinators by mimicking signals of food, mates or breeding sites without providing a reward to the pollinators. Although deceptive systems are generally considered ineffective at attracting pollinators compared with rewarding systems (Jersáková et al. 2006), the use of animal communities that are not otherwise involved in pollination may be advantageous because it allows escape from the ecological constraint of pollinator availability in time and space, as discussed in the context of the evolution of abiotic pollination (Cox 1991, Culley et al. 2002). Moreover, it may reduce the likelihood of heterospecific pollen transfer, as plants may trade pollinator visitation frequency for specificity (Scopece et al. 2010). Mycophagous insects are rarely involved in pollination but are abundant in shady forest floor ecosystems, which in turn are relatively devoid of flower-visiting insects (Policha et al. 2016). Nevertheless, few examples of pollination via mushroom mimicry have been reported (Vogel 1978, Johnson and Schiestl 2016, Suetsugu 2018), and the well-established example of pollination by mycophagous insects via mushroom mimicry is limited to the orchid genus Dracula, in which mycophagous drosophilids are attracted to flowers that morphologically and chemically resemble mushrooms (Endara et al. 2010, Policha et al. 2016, 2019). The rare occurrence of mushroom mimicry as a pollination system may indicate that the conditions under which the strategy can evolve are limited. For example, it may be because targeting specific groups of insects as pollinators is difficult, as most mycophagous insects have broad host ranges owing to the ephemeral nature of the fungi they feed on (Lacy 1984, Hanski 1989, but see Jonsell and Nordlander 2004). Moreover, demonstrating mushroom mimicry is often challenging, given that the floral traits associated with mushroom mimicry are not well understood, and the life histories of potentially mycophagous insects are largely unexplored. In the present study, we explored an unverified example of specialized pollination via mushroom mimicry in an aroid species, Arisaema sikokianum Franch. et Sav., although its striking morphological similarities to mushrooms have long been recognized (see Figure 1a) (Vogel and Martens, 2000). Arisaema (Araceae) consists of 200 spp., with its center of diversity in E. Asia (Murata et al. 2018). The genus is well known for its ‘one-way traffic’ mechanism, with most members being sequential hermaphrodites (Vogel and Martens, 2000), in which the spathe functions as a trap for the pollinators (typically dipterans). Only male inflorescences have an exit hole at the bottom of the spathe, and pollinators are trapped in the spathes of the female inflorescence, which does not have an exit, after its pollination. Because mushroom-associated insects such as fungus gnats pollinate some Arisaema species, mushroom mimicry has been suggested as their attraction mechanism (Vogel, 1978; Vogel and Martens, 2000). However, such mimicry has not been demonstrated conclusively, and most species in the genus do not have morphological or chemical characteristics shared with mushrooms. Here, we show that although not all Arisaema species have ‘mushroom flowers’, A. sikokianum has specialized floral traits for mushroom mimicry. To test the ‘mushroom flower’ hypothesis in A. sikokianum, we surveyed the pollinators trapped within the inflorescences and compared them with the insects attracted to various mushrooms. We further conducted comparative chemical analyses of floral and fungal volatile compounds to determine whether bioRxiv preprint doi: https://doi.org/10.1101/819136; this version posted October 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. A. sikokianum imitates a specific fungus. Finally, we investigated whether the mushroom mimicry in A. sikokianum results in a unique pollination mode that enables reproductive isolation from other sympatric Arisaema species. Materials and Methods Target plants Arisaema sikokianum is a tuberous perennial species endemic to western Japan and is ranked as vulnerable in the Red List of Japan (Ministry of the Environment, Government of Japan, 2012). This species is fairly rare across most of its distribution range and is ranked as highly endangered in most local (prefectural) red lists. However, the species is relatively abundant in Tokushima and Kochi prefectures. In mid to late April, the plant bears an inflorescence with a white, balloon-like appendix surrounded by a spathe that is white on the inside and dark purple on the outside (Figure 1a). As in most Arisaema spp., the inflorescence of A. sikokianum is unisexual, and the plant changes sex according to tuber size; larger tubers bear female inflorescences and smaller tubers male inflorescences (i.e., sequential hermaphroditism; Fukai, 2004). Male and female plants have almost identical inflorescences, except that a slit opening is present at the bottom of the inflorescence only in males, and the arrangement of the inflorescence tends to be higher in females. Arisaema japonicum Blume is closely related to A. sikokianum and belongs to the same section (Pistillata) (Ohi-Toma et al. 2016). The distribution ranges of A. japonicum and A. sikokianum overlap (Murata et al. 2018), and the two species sometimes exist sympatrically and bloom almost simultaneously. Arisaema japonicum has a similar growth habitat and ecology to A. sikokianum but is strikingly different in its floral characters, in that the inflorescence has a green, rod-shaped appendix surrounded by a green spathe (Figure 1b). Insect collection We field-collected